REPRO-HPC

1ˢᵗ Workshop in Sustainable Practices for Reproducibility in HPC

hosted at ISC26 - June 26th 2026

HomeCall for ContributionsProgramAccepted TalksCommittees

Reproducibility is a cornerstone for trustworthy and robust scientific progress.

The High-Performance Computing (HPC) community often faces reproducibility challenges due to complex software stacks, cutting-edge hardware, and costly operations (computations, data transfers, etc.). The reproducibility challenges could also be explained by a lack of education on what reproducibility is, a lack of tools offered by the platforms to support reproducibility, inconsistencies in the various venues’ guidelines for packaging the artifacts, or a lack of incentives for the authors to make an extra effort.

Overall, reproducibility in HPC is mostly a methodological and technical problem, which can only be addressed by gathering the community and discussing all together about the way forward.

This workshop brings together the HPC community (researchers, practitioners, platform providers, and educators) to share their feedback, tools, and best practices to tackle the reproducibility hurdles met in HPC.

Workshop Topics

The topics of interest of the workshop include, but are not limited to, the following:

General

  • Feedback/Lessons learned/Success stories from artifact authors, reviewers, and chairs, trying to package an experiment or trying to reproduce an experiment
  • Methods to create a “minimal reproducible experiment” to proxy the reproduction on the energy consuming full-scale version
  • Energy-efficient artifact reproduction in HPC
  • Case studies of sustainable (or unsustainable) artifact evaluation in HPC
  • Long-term reproducibility: ensuring artifacts remain accessible and evaluable as hardware/software evolves
  • Feedback from teaching HPC reproducibility principles
  • Reproducibility in the age of AI: concerns and opportunities

Software Environment / Workflow / CI/CD

  • Methods and tools to create a standalone and portable experiments (package managers, containers,…)
  • Methods and tools to support reproducibility in HPC from the early development stages (CI/CD, provenance, SBOM, …)
  • Methods and tools to support FAIR principles

Platforms (HPC Centers and Testbeds)

  • Tools and services that should be offered by HPC centers and testbeds to improve/support reproducibility
  • Billing, access, and “security” to HPC centers and testbeds for reproduction attempts

Artifact Evaluation Process

  • Proposals for new Artifact Evaluation processes (timelines, badges, reports, interactions between authors and reviewers, reviewer roles, etc.)
  • Incentives and recognition for Reproducibility in HPC (for authors and reviewers)
  • Human-centric sustainability: reducing reviewer/author fatigue and chair workload
  • Community standards for balancing rigor, efficiency, and human effort
  • “Proper” evaluation of proprietary software/hardware

Invited Speakers

Kate Keahey

Kate Keahey

ANL / UChicago

Title TBA

Helena Vela Beltran

Helena Vela Beltran

Do IT Now

EESSI: Addressing HPC Reproducibility Hurdles Through a Unified Software Stack

Program

Friday June 26th 2026:

Time (CET)Event
09:00 - 09:05Welcome and Introductions
09:05 - 09:40Invited Talk : Kate Keahey
09:40 - 09:55“Layered Reproducibility for High-Performance Computing Applications: The Feel++ and Ktirio Urban Buildings Case Study”
– J. Cladella, V. Chabannes, C. Prud’homme
09:55 - 10:10“iReVal: A Hardware-Aware LLM Agent for Artifact Reproducibility and Evaluation in HPC Systems”
– I. Benlamari
10:10 - 10:25“Standardising HPC Workflows for Sustainable Reproducibility: The JUBE Configuration Artefact Approach”
– J-O Mirus, F. S. M. Guimaraes, A. Sankaran, M. G. Barrios Sazo, C. Himmels, T. Breuer
10:25 - 11:00Invited Talk : Helena Vela Beltran – “EESSI: Addressing HPC Reproducibility Hurdles Through a Unified Software Stack”

Reproducibility is a cornerstone of trustworthy scientific progress, yet the high-performance computing (HPC) community frequently faces severe reproducibility challenges driven by complex software stacks, rapidly evolving hardware, and a historic lack of standardized deployment tools. To address these technical and methodological hurdles, this session introduces the European Environment for Scientific Software Installation (EESSI), a collaborative initiative designed to streamline how scientific software is deployed, shared, and reproduced across diverse infrastructures, including HPC clusters, cloud platforms, and local workstations. By providing a fully pre-built, module-based, and automated software environment, EESSI eliminates the inconsistencies often found across different platforms, offering a practical tool that helps researchers seamlessly package and replicate their computational environments.
11:00 - 11:30ISC Coffee Break
11:30 - 11:45“Nix to the Rescue for a Reproducible HPC-AI Software Stack”
– W. Du, J-M. Gratien, R. Gayno, B. Raffin
11:45 - 12:00“Some Lessons learnt on FPM promises with NixOS-Compose”
– A. Salmane, Y. Sun, H. Brunie, O. Richard
12:00 - 13:00Panel Discussion

Panelists

Accepted Talks

TitleAuthors
Nix to the Rescue for a Reproducible HPC-AI Software Stack Paper SlidesWenke Du; Jean-Marc Gratien; Raphael Gayno and Bruno Raffin
Standardising HPC Workflows for Sustainable Reproducibility: The JUBE Configuration Artefact Approach Paper SlidesJan-Oliver Mirus; Filipe Souza Mendes Guimaraes; Aravind Sankaran; Maria Guadalupe Barrios Sazo; Carina Himmels and Thomas Breuer
Some Lessons learnt on FPM promises with NixOS-Compose Paper SlidesAmine Salmane; Yifei Sun; Hugo Brunie and Olivier Richard
iReVal: A Hardware-Aware LLM Agent for Artifact Reproducibility and Evaluation in HPC Systems Paper SlidesInsaf Benlamari
Layered Reproducibility for High-Performance Computing Applications: The Feel++ and Ktirio Urban Buildings Case Study Paper SlidesJavier Cladella; Vincent Chabannes and Christophe Prud'homme

Call for Contributions

Submission Format

  • Submissions should be either a 2-page abstract or a 4-page short paper, excluding references, in the PDF format using the IEEE double column template

  • Accepted submissions will not appear in the ISC26’s proceedings, but will be published, alongside the presented slides, on the workshop’s webpage, unless explicit opt-out from the authors.

  • Accepted submission will have a 15-minute timeslot including presentation and Q&A.

  • Submission website: https://easychair.org/conferences/?conf=reprohpc26

Important Dates

  • Call for Contributed Talks open: March 2nd 2026

  • Submissions due: April 17th 2026 May 1st 2026 (AOE) May 6th 2026 (EoB) (final extension)

  • Notifications sent: May 8th 2026 May 20th 2026

  • Program finalized: May 26th 2026

  • Camera Ready version due: June 3rd 2026

  • Workshop Date: June 26th 2026

Committees

Organizing Committee

Quentin Guilloteau

Quentin Guilloteau

INRIA

Web Chair

Valérie Hayot-Sasson

Valérie Hayot-Sasson

ETS Montréal

Publicity Chair

Dennis Hoppe

Dennis Hoppe

HLRS

Josef Weidendorfer

Josef Weidendorfer

Dresden University of Technology

Program Chair

For any questions, please reach us at : repro-hpc@groupes.renater.fr

Abstract Reviewing Committee

  • André Bauer, Illinois Institute of Technology
  • Insaf Benlamari, Leibniz Supercomputing Center
  • Raphaël Bleuse, Univ. Grenoble Alpes
  • Tainā Coleman, San Diego Supercomputer Center
  • Ludovic Courtès, INRIA
  • Abdullatif Eymash, HLRS
  • Maxime Gonthier, INRIA
  • Samuel Grayson, University of Illinois at Urbana-Champaign
  • Arjun Parab, Leibniz Supercomputing Centre
  • Millian Poquet, Univ. Toulouse III
  • Bruno Raffin, INRIA
  • Amir Raoofy, Leibniz Supercomputing Centre